[net.news.sa] UUCP map files - What's in a name?!?

pleasant@caip.RUTGERS.EDU (Mel Pleasant) (08/28/86)

Recently I've received several questions which pertain to the naming
of UUCP map files posted in mod.map and how these names relate to the
contents of the files themselves.

Under the new naming scheme, a map file name begins with u. (the
prefix), and ends with \.[1-9][0-9]* (the suffix).  What falls between
the prefix and the suffix (the root) represents some logical grouping
of sites.  In general, a logical grouping will have some relation to a
geographical region of the globe, but this may not always be the case.
The prefix, which is always the same, tells you that the file is
indeed a map file.  The suffix of the first file of a logical grouping
starts with the number 1 and is incremented as needed as files within
the group are broken into smaller parts.  Thus, the suffixes tell you
how many files there are representing each of the logical groupings.
Other than what has just been stated here about map file names,
nothing else is guaranteed or implied.

When the new naming scheme was put into place, it was decided that a
file should not contain more than 64k characters.  Certain mailers and
older versions of the notes software can not handle messages larger
than 64k.  When a file, through expansion, grows beyond this limit, it
is broken into one or more files by the person handling it.  He may do
this in any way that seems appropriate to him e.g. he may split map
entries between these files as he chooses.  He may take the contents
of two overgrown files and split them into 3 new files, causing map
entries that were in file u.x.1 to suddenly appear in file u.x.2 (as
well as map entries in file u.x.2 to appear in file u.x.3).  He may
even reorganize the entries in an entirely new fashion.  In other
words, there isn't any guarantee that a map entry will appear in the
same file from month to month.  We only guarantee that map entries
that fall within a logical group will be grouped together in files
whose names share the same root, that the files will not be larger
than 64k and that the contents will be properly formatted.

In a practical sense of course, we've gone further than what has been
described above.  For instance, the 2nd part of a map file name is the
3-character ISO code representing the country that the site resides
in.  In the USA, there are further divisions of the name representing
which state the site resides in as well.  Sometimes a map file name
may contain some information pertaining to an organization.  The thing
to remember is that software written to read these files should not
depend upon these further divisions being the same from month to
month.  At most, the software should only depend upon the concept that
files with like roots represent a logical grouping of one form or
another.  Sometimes new roots may appear while others may be removed.
In general, we will attempt to announce changes as early as possible.
However, the naming scheme was designed to allow us to make changes to
filenames "on the fly".  If we now begin to guarantee more than what
has been described we will lose this flexibility to make changes as
the needs arise.
-- 

-Mel Pleasant
uucp:   ...{harvard, seismo, ut-sally, sri-iu, ihnp4!packard}!topaz!pleasant
arpa:   PLEASANT@RUTGERS